The general goal of the analysis is to understand the effect of the Canadian Registered Retirement Saving Plan (RRSP) on the financial status of the elderly in Canada and to compare the Canadian RRSP experience with IRA and 401(k) experience in the United States. The analysis should improve general understanding of how such programs can affect saving behavior and the well-being of the elderly. The Canadian experience affords a unique opportunity to evaluate the effects of a long standing program that is strutted very much like Individual Retirement Accounts in the United States. The Canadian data are ideally suited to evaluated the net saving effect of the RRSP program in a way that is easily compared with the U.S. IRA experience. In addition, the Canadian data offer two further opportunities. Because there is no penalty for withdrawal of RRSP funds before retirement, it is natural to evaluate the extent to which contributions are "locked in" until retirement. The data also allow evaluation of the saving effect of the new carry-forward provision, the effects of which are not easily predicted based on standard financial models of retirement behavior. To this end, there are six specific aims: (1) To determine from micro data how individuals saving in other forms changed as RRSP saving increased in the period since 1969. (2) To determine the extent to which RRSP saving is dedicated to retirement, even though in Canada there is no penalty for withdrawal of funds before retirement. (3) To estimate a formal model aimed at determining the net saving effect of the RRSP plan, to consider out-of-sample tests of the model's predicative validity, and to test the formal model results against alternative non-parametric estimates of the degree of substitution between RRSP and other forms of financial assets saving. (4) To understand the effect on saving of the new RRSP carry- forward provision that was introduced in 1992. (5) To project contribution rates forward to determine the effect of the RRSP program on the wealth of the elderly. (6) To contribute more generally to our understanding of the determinants of saving behavior.